Is there any way to do the following in SPSS? I have a set of 12 stimulus materials that I have to give to each of my subjects in random order (I have 60 subjects). In order to randomize the stimuls materials, I wanted to have SPSS generate a variable that could shuffle the numbers 1-12 in random order, i.e., sample WITHOUT replacement. If the randomization could be done with replacement, it would be no problem, it would simply be a matter of using the uniform distribution: COMPUTE X= TRUNC(RV.UNIFORM(1,13)). Is there anyway to sample WITHOUT replacement? Sylvia *****************************************************. *Sylvia, Rich and SPSSX-Listers, *The following solves the permutation problem concisely. *It illustrates some general SPSS approaches for. 1: turning one case into multiple cases 2: Generating random permutations 3: turning multiple cases into one case 4: Preserving ones mouse, keyboard...sanity ;-), (These issues arise frequently). *Regards, David Marso *SPSS Consulting Services ----------------------------------------- * Here are the stimuli *. DATA LIST / STIM01 TO STIM13 1-13 (A). BEGIN DATA ABCDEFGHIJKLM END DATA. * Create 100 copies (one per subject) *. VECTOR STIM=STIM01 TO STIM13. STRING STIMX(A1). LOOP ID=1 TO 100. + LOOP STIMNUM=1 TO 13. + COMPUTE STIMX=STIM(STIMNUM). + XSAVE OUTFILE 'TMP' / KEEP ID STIMX. + END LOOP. END LOOP. EXE. GET FILE 'TMP'. * Permute the stimuli * . COMPUTE JUMBLE=UNIFORM(1). SORT CASES BY ID JUMBLE. RANK VARIABLES= JUMBLE BY ID . * Get the stimuli onto one line per subject *. STRING STIM01 TO STIM13 (A1). VECTOR STIM = STIM01 TO STIM13. COMPUTE STIM(RJUMBLE)=STIMX. AGGREGATE OUTFILE * / BREAK ID / STIM01 TO STIM13 = MAX(STIM01 TO STIM13).